Tutorial
Wednesday December 16, 2015   

Deep Learning for Computer Vision
Prof. C. V. Jawahar
IIIT Hyderabad

Abstract: There has been an increasing interest in use of deep learning for computer vision ever since the success of AlexNet in 2012. This tutorial aims at providing an overview of the recent advances in this area. This tutorial is aimed at research students and beginner researchers. We provide an overview of the popular deep learning techniques, and validate their application with a set of case studies from the recent years. Some amount of exposure to the modern computer vision and machine learning is expected. Tutorial is broadly divided into broadly four parts as below:


Part 1: Introduction
          - Computer vision: brief history and problem space
          - Neural network learning

Part 2: Introduction to Deep learning
          - Architectures
          - Algorithms

Part 3: Recent success
          - Case studies
          - Discussions

Part 4: Various
          - Practical aspects
          - Challenges

Professor C. V. Jawahar

Biography: Dr. C.V. Jawahar, has been the key instrumental person in motivating for Research. He has also played a role in defining and directing the research in International Institute of Information Technology, Hyderabad which was established in 1998. Dr. C.V. Jawahar has the following Research Areas:  Indian Language Document Image Understanding, Geometric Approach to Computer Vision Problems, Content Based Recognition and Retrieval from Multimedia Data, Machine Learning and Pattern Recognition. He has published more than 300 national and international highly cited papers.

 


Tutorial
Wednesday December 16, 2015

3D Reconstruction from Images and Videos


Prof. Subhashis Banerjee
Dept. of Computer Science and Engineering,
IIT Delhi

Dr. Venu Madhav Govindu
Assistant Prof. Dept. of Electrical Engineering
IISc Bangalore

Abstract: In this tutorial talk we will cover techniques from computer vision and algorithms for 3D reconstruction, mapping and model creation from a collection of community acquired images and videos. To this end we will cover the basics of projective geometry as applied to computer vision and discuss the issues of camera auto-calibration, image and video features, match graphs, motion averaging and optimization techniques for bundle adjustment and model creation. We will also discuss some recent techniques for Simultaneous Localization And Mapping (SLAM) from community acquired videos, as can be obtained using phone cameras. These techniques can have wide application in robotics, autonomous navigation etc.
In addition, we will cover some photometric techniques for obtaining dense 3D reconstructions using inexpensive and off the shelf image and depth acquisition devices. In this context we will also discuss some techniques for registration and manipulation of 3D point clouds.
We will demonstrate some results of 3D reconstruction using each of the above ideas.
Target audience: Graduate students and young researchers with interest in computer vision and computer graphics.
Recommended background: Familiarity with linear algebra and basic optimization techniques
Tutorial topics:

  • Basics of projective geometry
  • Camera calibration and self-calibration
  • Multiple views geometry
  • Automated large scale model creation from community acquired photographs
  • Motion averaging
  • Hierarchical large scale 3D reconstruction.
  • Dense reconstruction using photometric stereo
  • Point cloud registration
  • Simultaneous localization and mapping

 

Professor Subhashis Banerjee

Biography: Subhashis Banerjee is a professor at the department of Computer Science and Engineering, Indian Institute of Technology, Delhi, India. His broad areas of research include Computer Vision, Robotics, Real-time systems, Image processing and Pattern recognition.  He is on the editorial boards of International Journal of Computer Vision and Computer and Graphics.

Professor Venu Madhav Govindu

Biography: Venu Madhav Govindu received the Ph.D. degree in electrical engineering from the University of Maryland, College Park, MD, USA. He is currently an Assistant Professor with the Department of Electrical Engineering, Indian Institute of Science, Bengaluru, India. His current research interests include geometric and statistical inference problems in computer vis